CN105630161A - Android system-based gesture control ring and using method thereof - Google Patents

Abstract

The invention provides an Android system-based gesture control ring and a using method thereof. The invention relates to an Android system-based intelligent ring. The intelligent ring consists of an existing three-axis acceleration sensor, a microcontroller, a Bluetooth module and a power supply. After being connected with a mobile phone, the intelligent ring is capable of carrying out processing in various odes through predefined gestures, achieving the effect of carrying out multiple controls through storing few gestures through defining meta gestures and complicated gestures, changing the current mobile phone control effect through changing the mobile phone function interface mapping relationship, and realizing the personalized control gestures of the users through learning software set gestures; and according to the Android system-based gesture control ring and the using method thereof, barrier-free recognition can be carried out, user-friendliness can be provided, remote operation can be realized indoors without being limited by the direction and position, and control can be realized in a Bluetooth communication range.

Description

Gesture based on android system controls finger ring and using method thereof

Technical field

The present invention relates to a kind of intelligence finger ring, especially a kind of intelligent finger ring based on android system, further relate to the using method of a kind of intelligent finger ring based on android system.

Background technology

In recent years, along with the development of electronic technology, many sensors are more and more integrated, and power consumption constantly reduces, and performance improves constantly, and most importantly volume is also in continuous decline, and this also makes intelligence wearable device constantly be pursued. In traditional smart machine platform, ripe man-machine interaction mode (referring mainly to mobile phone here) is mainly based on mechanical keyboard, touch screen and phonetic entry, and gesture is a part indispensable in daily life as body language, if gesture and Intelligent worn device are combined, become the input scheme under some scenes, it will improve people greatly and control the efficiency of mobile phone.

Such as, but forget mobile phone when using mobile phone and be placed in where; Or but want when both hands are busy to receive calls; Or suddenly enter into the environment thinking to see clearly rapidly surrounding in dark surrounds; Also or think that when facing a danger a key is reported to the police or recording etc. For mobile phone itself, it is desirable to realizing purpose in above-mentioned scene needs troublesome operation, as the first scene, need to find an other mobile phone dialing phone or the memory with oneself to go for, expending time in very much, if called with voice, then cannot environment noisy again realize again; Or for the third scene, when facing a danger, it is likely to not take out the chance of mobile phone at all, now by gesture operation, directly invokes the correlation function of mobile phone, such as flash lamp or phone contacts, then can solve problem in the very first time. And for existing commercially available wearable product, function is largely focused on the aspects such as note step, sleep monitor, prompting message, identification, to provide the user with information, almost have ignored the inconvenience that under above-mentioned scene, mobile phone uses even completely.

Then, design and a meet people's regular job experience, the control to Android smartphone (hereinafter referred to as mobile phone) can be strengthened and experience, and the finger ring that controls of inconvenience situation in similar scene can be solved just become our research direction.

Summary of the invention

In order to overcome the deficiencies in the prior art, the present invention devises a kind of Intelligent gesture based on android system and controls finger ring, hereinafter referred to as finger ring, aim to solve the problem that in daily mobile phone operation, the traditional operation modes such as touch, voice and button are relied on merely to control the limitation of mobile phone, fill up and utilized finger ring collection acceleration information to be processed by data, the final vacancy controlling mobile phone, enhance the experience that mobile phone controls simultaneously, under many scenes can convenient, promptly calling mobile phone function, reach the purpose of man-machine interaction more freely.

The technical solution adopted for the present invention to solve the technical problems is:

The Intelligent gesture based on android system designed by the present invention controls finger ring and comprises a 3-axis acceleration sensor, a microcontroller, a bluetooth module and a power supply, wherein 3-axis acceleration sensor is by SPI interface and microcontroller communication, microprocessor is by bluetooth module and Android intelligent communication, and microcontroller, 3-axis acceleration sensor and bluetooth module are powered by the power supply that one piece of voltage is 3.7 volts.

The present invention also provides for the using method relating to controlling finger ring based on the Intelligent gesture of android system, and step is as follows:

Step 1: predefined

1.1 Pre-defined gestures are divided into unit's gesture and combination gesture, unit's gesture is to make finger ring move according to a direction, ensure that three axles of its built-in 3-axis acceleration sensor at most have two axles to produce acceleration information, such gesture is minimum inseparable recognition unit, and the combination of multiple units gesture is defined as complicated gesture;

1.2 Pre-defined gesture mode of operations, have gesture learning model, gesture strengthening cooling, gesture match pattern respectively:

Gesture learning model refers to the learning model of gesture, forms characteristic vector by current acceleration information, be called template after over-sampling, feature extraction;

Gesture strengthening cooling finger ring constantly trains template according to the gesture motion of user, namely the acceleration information by the front acceleration information once storing gesture and current gesture are collected is averaged, and as corresponding new gesture, meansigma methods is replaced original template;

The normal mode of operation that namely template matching model is given tacit consent to, finger ring to collecting current acceleration information, first feature extraction, again through identifying current signature vector, match the template of storage, calculate the control command that template is corresponding, it is achieved the associative operation of mobile phone terminal;

Step 2: after finger ring powers on, it is judged that whether the match is successful with mobile phone for finger ring, if unsuccessful, then the standby mode entering low-power consumption waits that the match is successful; If success, then finger ring entrance step 3 carries out the multiple-limb selection of gesture mode of operation;

Step 3: if selecting gesture learning model, then enter step 4, if selecting gesture strengthening cooling, then enter step 5, if selecting gesture match pattern, then enter step 6, before entering every kind of gesture mode of operation, acceleration information is carried out the pretreatment operation of hop removal, mean filter, feature extraction;

Step 4: if selecting gesture learning model, after the acceleration information received is carried out the pretreatment in step 3 by microcontroller, storage order and characteristic vector, the gesture mode of operation returning step 3 after terminating selects state;

Step 5: if selecting gesture strengthening cooling, current gesture acceleration information is carried out the pretreatment in step 3 by microcontroller, average with the acceleration information of last gesture again, and average result is replaced last characteristic vector as new characteristic vector, the gesture mode of operation returning step 3 after terminating selects state;

Step 6: if selecting gesture match pattern, after acceleration information is carried out the pretreatment described in step 3 by microcontroller, start coupling from first characteristic vector of storage template, by dynamic time warping algorithm DTW (DynamicTimeWarping) carry out asking from:

(1) when range estimation is not up to threshold value, then order characteristic vector coupling rearwards;

(2) if range estimation meets or exceeds threshold value, then it is considered as this characteristic vector of successful match, thereafter order corresponding for characteristic vector is sent to mobile phone by bluetooth module, send the gesture mode of operation returning step 3 after successfully and select state to continue to select;

If it fails to match, resume waiting for new gesture and carry out matching operation, or user selects to enter step 4 voluntarily, re-starts gesture Template Learning.

The invention has the beneficial effects as follows: by defining unit's gesture, form complicated gesture again through being mutually combined of unit's gesture, then reach to store a small amount of gesture and can reach the effect of various control; Gesture correspondence cell-phone function is variable, namely by changing cell-phone function interface mappings relation, can change the current effect that mobile phone is controlled; Learnt by the gesture of software design patterns, user can self-defined unit gesture, thus realizing the control hands gesture of user personality; Accessible identification, user friendly, long-range operation can be realized in indoor, not by direction and position limitation, as long as within the scope of bluetooth communication, manipulation can be realized.

Accompanying drawing explanation

The finger ring system that Fig. 1 is the present invention runs entire block diagram.

Fig. 2 is the software flow entire block diagram of the present invention.

Detailed description of the invention

Below in conjunction with drawings and Examples, the present invention is further described.

The Intelligent gesture based on android system designed by the present invention controls finger ring and comprises a 3-axis acceleration sensor, a microcontroller, a bluetooth module and a power supply, wherein 3-axis acceleration sensor is by SPI interface and microcontroller communication, microprocessor is by bluetooth module and Android intelligent communication, and microcontroller, 3-axis acceleration sensor and bluetooth module are powered by the power supply that one piece of voltage is 3.7 volts.

The hardware composition of this design has: ADXL345 acceleration sensor module, with the STM32 microcontroller being core, power module, bluetooth module and intelligence Android mobile phone terminal. After hardware powers on, first the Bluetooth pairing carrying out finger ring and mobile phone connects, after control module judges successful connection, start the reading to acceleration transducer data and judgement, otherwise finger ring enters low-power consumption battery saving mode, waits to be connected waking up, after successful connection, start recording acceleration data, enter normal duty.

Specifically comprising the following steps that of using method

Step 1: predefined

1.1 Pre-defined gestures, are divided into unit's gesture and combination gesture, and unit's gesture is to make finger ring move according to a direction, it is ensured that three axles of its built-in acceleration sensor at most have two axles to produce acceleration information, and such gesture is minimum inseparable recognition unit; The gesture combination of multiple units is defined as complicated gesture, is combined as appropriate value scope less than first gesture of 4 kinds of first gestures.

1.2 Pre-defined gesture mode of operations, are gesture learning model, gesture strengthening cooling, gesture match pattern respectively:

Gesture learning model refers to the learning model of gesture, forms characteristic vector by current acceleration information, be called template after over-sampling, feature extraction;

Gesture strengthening cooling finger ring constantly trains template according to the gesture motion of user, namely the acceleration information by the front acceleration information once storing gesture and current gesture are collected is averaged, and as corresponding new gesture, meansigma methods is replaced original template;

The normal mode of operation that namely template matching model is given tacit consent to, finger ring to collecting current acceleration information, first feature extraction, again through identifying current signature vector, match the template of storage, calculate the control command that template is corresponding, it is achieved the associative operation of mobile phone terminal;

Step 2: after finger ring powers on, it is judged that whether the match is successful with mobile phone for finger ring, if unsuccessful, then the standby mode entering low-power consumption waits that the match is successful; If success, then finger ring entrance step 3 carries out the multiple-limb selection of gesture mode of operation;

Step 3: if selecting gesture learning model, then enter step 4, if selecting gesture strengthening cooling, then enter step 5, if selecting gesture match pattern, then enter step 6, before entering every kind of gesture mode of operation, acceleration information is carried out the pretreatment operation of hop removal, mean filter, feature extraction;

Described feature extraction, is first the intercepting of acceleration information, it is stipulated that starting and ending is all realize by the duration steady gesture no less than 0.5 second, namely when this gesture occurs and keeps more than 0.5 second, being considered as gesture operation to start or terminate, now, the Z axis of acceleration transducer is vertically upward, positive and negative deviation is less than 0.2g, g is acceleration of gravity, and X, Y-axis are zero, and positive and negative deviation is less than 0.2g,, it is judged that threshold interval condition is:

$\left\{\begin{array}{c}0.8<\frac{1}{N}\underset{k=i}{\overset{N}{\&Sigma;}}{Z}_i<1.2\\ -0.2<\frac{1}{N}\underset{k=i}{\overset{N}{\&Sigma;}}{X}_i<0.2\\ -0.2<\frac{1}{N}\underset{k=i}{\overset{N}{\&Sigma;}}{Y}_i<0.2\end{array}\right.$

Wherein, N is counting of collecting of acceleration transducer, X in 0.5s_i��Y_i��Z_iFor the acceleration information that acceleration transducer collects at tri-axles of i-th moment X, Y, Z, the value of k is i, i+1, i+2 ..., N, after meeting conditions above, start gesture operation simultaneously, now intercept out the acceleration information started between two the steady gestures terminated, the acceleration information collected between the recompression steady gesture of the two, specifically chosen formula is:

$a_{t_k}=a_{t_1}+(t_k-t_1)\&times;\frac{a_{t_2}-a_{t_1}}{t_2-t_1}$

Wherein, t_kFor the kth moment currently calculated,For the acceleration information of X-axis corresponding to tk moment, Y-axis or Z axis, t₁��t₂Two respectively adjacent moment,Respectively t₁��t₂The X-axis at some place, Y-axis or Z axis acceleration information, remove t₁��t₂Middle data, only retain t₁��t₂The position of point and data value information, i.e. t_kThe data of point can be usedRepresent, thus realizing the extraction of feature.

Step 4: if selecting gesture learning model, after the acceleration information received is carried out the pretreatment in step 3 by microcontroller, storage order and characteristic vector, the gesture mode of operation returning step 3 after terminating selects state;

Step 5: if selecting gesture strengthening cooling, current gesture acceleration information can be carried out the pretreatment in step 3 by microcontroller, average with the acceleration information of last gesture again, and meansigma methods is replaced last characteristic vector as new characteristic vector, the gesture mode of operation returning step 3 after terminating selects state;

Step 6: if selecting gesture match pattern, after acceleration information can be carried out the pretreatment described in step 3 by microcontroller, first characteristic vector from memory space starts coupling, by dynamic time warping algorithm DTW (DynamicTimeWarping) carry out asking from, threshold value selects T/2��t��2T, wherein T is the time used by template gesture, and t is the time used by current gesture, then:

When range estimation is not up to threshold value, then order characteristic vector coupling rearwards;

If range estimation meets or exceeds threshold value, then it is considered as this characteristic vector of successful match, thereafter order corresponding for characteristic vector is sent to mobile phone by bluetooth, send the gesture mode of operation returning step 3 after successfully and select state.

The model selection of step 3, acquiescence is the match pattern that gesture identification controls mobile phone, can enter into data prediction after collecting the data, including hop removal, mean filter, feature extraction, thus calculating corresponding mobile phone control command. Being directed to different mode, the direction after feature extraction is also different, under learning model, can the feature extracted be stored as template; And under strengthening cooling, meeting constantly weighting also updates current template, final discrimination is made to be greatly promoted.

Claims (2)

1. control a finger ring based on the gesture of android system, including a 3-axis acceleration sensor, a microcontroller, a bluetooth module and a power supply, it is characterised in that:

3-axis acceleration sensor is by SPI interface and microcontroller communication, and microprocessor is by bluetooth module and Android intelligent communication, and microcontroller, 3-axis acceleration sensor and bluetooth module are powered by the power supply that one piece of voltage is 3.7 volts.

2. one kind utilizes the using method controlling finger ring described in claim 1 based on the gesture of android system, it is characterised in that comprise the steps:

Step 1: predefined

1.1 Pre-defined gestures are divided into unit's gesture and combination gesture, unit's gesture is to make finger ring move according to a direction, ensure that three axles of its built-in 3-axis acceleration sensor at most have two axles to produce acceleration information, such gesture is minimum inseparable recognition unit, and the combination of multiple units gesture is defined as complicated gesture;

1.2 Pre-defined gesture mode of operations, have gesture learning model, gesture strengthening cooling, gesture match pattern respectively:

Gesture learning model refers to the learning model of gesture, forms characteristic vector by current acceleration information, be called template after over-sampling, feature extraction;

Gesture strengthening cooling finger ring constantly trains template according to the gesture motion of user, namely the acceleration information by the front acceleration information once storing gesture and current gesture are collected is averaged, and as corresponding new gesture, meansigma methods is replaced original template;

The normal mode of operation that namely template matching model is given tacit consent to, finger ring to collecting current acceleration information, first feature extraction, again through identifying current signature vector, match the template of storage, calculate the control command that template is corresponding, it is achieved the associative operation of mobile phone terminal;

Step 2: after finger ring powers on, it is judged that whether the match is successful with mobile phone for finger ring, if unsuccessful, then the standby mode entering low-power consumption waits that the match is successful; If success, then finger ring entrance step 3 carries out the multiple-limb selection of gesture mode of operation;

Step 3: if selecting gesture learning model, then enter step 4, if selecting gesture strengthening cooling, then enter step 5, if selecting gesture match pattern, then enter step 6, before entering every kind of gesture mode of operation, acceleration information is carried out the pretreatment operation of hop removal, mean filter, feature extraction;

Step 4: if selecting gesture learning model, after the acceleration information received is carried out the pretreatment in step 3 by microcontroller, storage order and characteristic vector, the gesture mode of operation returning step 3 after terminating selects state;

Step 5: if selecting gesture strengthening cooling, current gesture acceleration information is carried out the pretreatment in step 3 by microcontroller, average with the acceleration information of last gesture again, and average result is replaced last characteristic vector as new characteristic vector, the gesture mode of operation returning step 3 after terminating selects state;

Step 6: if selecting gesture match pattern, after acceleration information is carried out the pretreatment described in step 3 by microcontroller, start coupling from first characteristic vector of storage template, by dynamic time warping algorithm DTW (DynamicTimeWarping) carry out asking from:

(1) when range estimation is not up to threshold value, then order characteristic vector coupling rearwards;

(2) if range estimation meets or exceeds threshold value, then it is considered as this characteristic vector of successful match, thereafter order corresponding for characteristic vector is sent to mobile phone by bluetooth module, send the gesture mode of operation returning step 3 after successfully and select state to continue to select;

If it fails to match, resume waiting for new gesture and carry out matching operation, or user selects to enter step 4 voluntarily, re-starts gesture Template Learning.